English | EN 
“Innovation distinguishes between a leader and a follower.” – Steve Jobs
This is particularly resonant for product managers/owners who are constantly looking out for ideas and tools to innovate & disrupt quickly, and at scale. Adoption of Large Language Model (LLM) based tools is not just about embracing new technology — it’s a statement of vision and leadership in an industry ripe for transformation.
By integrating LLM tools like ChatGPT, Claude etc. into their day-to-day tasks, Product Managers (PM) & Product Owners (PO), are not merely managing products; they are leading the charge towards a more intuitive, efficient, and customer-centric experiences.
In part 1 of the series, we try to explore a few ways for PMs & POs to integrate Generative AI into their day-to-day responsibilities, using examples from Commercial Property Insurance but these tips can be easily applied to any industry:
Brainstorming
Idea Generation:
Begin by engaging LLMs in brainstorming session. Input the current market challenges, customer pain points, desires, and your strategic goals. Factor in innovation, technology, customer engagement and regulatory constraints.
Prompt example:
Our goal is to address challenges faced by large and mid-size customers in Commercial Property. Here is an outline of customer pain points, desires, and vision:
High insurance costs, complex policies & slow claims processing continues to be challenges in Commercial Property Insurance for customers, both big & small. Customers are constantly looking for more affordable, flexible insurance options that can be easily customized to their specific needs while demanding faster and transparent claims. Customers may desire impact to their policy because of regulation, and how they are protected against climate changes. The strategic vision in this case could be to create a seamless experience to the customer quoting or a claims experience.
Brainstorming considerations:
- What novel product features or packages can we implement to address insurer’s pain points?
- How can we leverage technologies like AI and Blockchain to simplify policy complexity, personalize products and speed up claims processing?
- What approaches will improve transparency, build broker trust, and improve customer experience?
- How do we build insurance products that are flexible and adaptive to regulatory & climate policies?
Prioritize Ideas:
Once you have a list of ideas, the next step is to prioritize them. One of my favorite frameworks is looking at ideas through the lens of four risks – business risk, usability risk, engineering risk, and viability risk. To take this further, you could ask your LLM to score each of your ideas on these four dimensions as a starting point and tweak it further with your team & stakeholders.
Prompt example:
Assess the proposed insurance product ideas in terms of Business Risk, Usability Risk,
Engineering Risk, and Viability Risk. Each idea should be scored quantitatively.
Ideas for Evaluation:
{Insert brainstormed ideas generated from previous section}
Score each idea using below risk factors, on a scale from 1 to 5, where 1 indicates the highest risk and 5 indicates the lowest risk:
Business Risk: Will customers use the product feature? Consider market demand and customer preferences.
Usability Risk: Will it be easy for customers to use? Evaluate user interface and experience. Engineering Risk: Can we build it easily? Assess technical feasibility and resource availability. Viability Risk: Is the product feature economically and operationally sustainable over time?
From Abstract Ideas to Concrete Features
Feature Definition:
Once you have your ideas prioritized, the next step is to pick the ideas you want to take to the build stage. At this point, you will need to refine the ideas into concrete features. Ask your LLM to outline a feature capturing business goal, values, description, high level acceptance criteria and technical consideration.
Prompt example:
For each of following ideas {Insert prioritized ideas from above}, write a feature using following template:
Feature Name: [Name of the Feature]
Business Goal: What is the primary objective this feature aims to achieve in the context of the insurance business?
Core Values: What are the key values this feature upholds (e.g., customer satisfaction, innovation, efficiency)?
Feature Description: Provide a detailed description of the feature. How does it function? What are its key components?
High-Level Acceptance Criteria: Outline the criteria that must be met for this feature to be considered complete and successful.
Technical Considerations: Identify any major technical challenges or requirements that need to be addressed in the development of this feature.
Stitching Features Together:
LLMs can envision how individual features can be integrated into a cohesive user experience. For instance, integrating a chatbot for instant claims reporting with backend automation for immediate claim processing. As another example you may be looking for way to create integrated experience for policy admin features and automated underwriting to evaluate how you can use algorithms instantly priced based on entered coverage information.
Fine-tuned LLMs like DesignerGPT can design webpages. While foundation models can generate HTML code, they tend to have drab UI layouts and inconsistent design elements. DesignerGPT is fine tuned to create cleaner UIs and generate HTML code that get automatically deployed on an Online IDE like Replit. As a note of caution, LLMs tend to “forget” previous components as you iterate through design and so you will have to “remind” to add back those use cases that got missed!
Hope you enjoyed this blog. Are you using LLMs to enhance your role as a PM? Share your experiences and insights below!
